Optimal Intereye Difference Thresholds in Retinal Nerve Fiber Layer Thickness for Predicting a Unilateral Optic Nerve Lesion in Multiple Sclerosis

Abstract

Background: The optic nerve is a frequent site for involvement in multiple sclerosis (MS). Optical coherence tomography (OCT) detects thinning of the retinal nerve fiber layer (RNFL) in eyes of patients with MS and in those meeting criteria for clinically or radiologically isolated demyelinating syndromes. Current international diagnostic criteria for MS do not include the optic nerve as an imaging lesion site despite the high prevalence of acute optic neuritis (ON), or occult optic neuropathy, among early MS and clinically isolated syndrome patients; as well as most MS patients over the course of the disease. We sought to determine optimal thresholds for intereye difference in peripapillary RNFL thickness that are most predictive of a unilateral optic nerve lesion.

Methods: We analyzed spectral domain OCT data of 31 healthy volunteers and 124 patients with MS at a single center as part of an ongoing collaborative investigation of visual outcomes. Intereye differences in peripapillary (360°) RNFL thickness were calculated as the absolute value of the difference. First, we determined the 95th percentile value of intereye difference for the healthy volunteers. This value was applied to the convenience sample group of MS patients as a validation cohort determining how well this threshold could distinguish patients with vs without a history of unilateral ON. The relation of intereye differences in peripapillary RNFL thickness to binocular low-contrast letter acuity scores was also examined.

Results: Among healthy volunteer participants (n = 31), the 95th percentile value for intereye difference (upper boundary of expected for normal controls) was 6.0 μm. This value was applied to the convenience sample group of MS patients (n = 124, validation cohort). Positive predictive value, negative predictive value, sensitivity, and specificity for identifying MS patients with a history of unilateral ON were calculated for the 6-μm threshold value in a 2 × 2 table analysis with the application of χ tests (P < 0.0001). The 6-μm threshold was predictive of worse binocular low-contrast acuity scores at 2.5% (P = 0.03) and 1.25% (P = 0.002 by linear regression analyses). A receiver operating characteristic curve analysis demonstrated an optimal intereye difference threshold of 5 μm for identifying unilateral ON in the MS cohort.

Conclusions: An intereye difference of 5-6 μm in RNFL thickness is a robust structural threshold for identifying the presence of a unilateral optic nerve lesion in MS.

FIG. 1.

Scatter plot demonstrating inter-eye differences in RNFL thickness for healthy controls and for patients with MS. Means are indicated by the red bars.

FIG. 2.

Receiver operating characteristic (ROC) curve examining inter-eye difference in RNFL thickness as a continuous variable for distinguishing patients with MS with vs. without a history of acute unilateral ON. The optimal threshold value at which sensitivity and specificity are closest together for inter-eye RNFL thickness difference is 5 microns; this value is represented by the point of inflection on the ROC curve indicated by the arrow. The area under the ROC curve is the probability that inter-eye difference in RNFL thickness will correctly classify MS patients with vs. without an acute unilateral ON history. ROC curve areas range between 0.5 (capacity to distinguish no better than flipping a fair coin) and 1.0 (perfect ability to distinguish).